The Cancer Modeling Section seeks to elucidate the complex molecular/genetic program governing tumor genesis and progression through the development and analysis of genetically engineered mouse models of human cancer. Our efforts in this regard are focused primarily on cutaneous malignant melanoma. Exposure to ultraviolet (UV) radiation is a causal agent in the vast majority of melanoma. Retrospective epidemiological data have suggested that melanoma is provoked by intermittent, intense exposure to UV, particularly during childhood. Previously, we tested this hypothesis in transgenic mice in which the receptor tyrosine kinase MET was deregulated by virtue of ectopic expression of its ligand, hepatocyte growth factor/scatter factor (HGF/SF). We discovered that a single neonatal dose of burning UV radiation in these mice was necessary and sufficient to induce tumors reminiscent of human melanoma with shortened latency (Noonan et al., Nature 413: 271-2, 2001). A critical role for the INK4A/ARF locus, which helps regulate the pRb and p53 pathways and is widely regarded as a key melanoma suppressor in human patients, was also confirmed in our animal model (Recio et al., Cancer Res. 62: 6724-30, 2002). There has been controversy surrounding the relative risks associated with UVB versus UVA radiation. We used albino HGF/SF transgenic mouse to show that UVB, but not UVA, alone is able to induce the full melanoma phenotype in the absence of pigment (DeFabo et al., Cancer Res. 64: 6372-6, 2004). However, we recently showed that UVA is highly melanomagenic in pigmented HGF/SF-transgenic mice (Noonan et al., Nat. Commun. 3: 884, 2012), demonstrating that melanin is associated with oxidative DNA damage and mutagenesis, and thus represents a double-edged sword with respect to melanoma risk. Our work also suggests that indoor tanning, which is mostly UVA-based, could be a significant health risk. The relevance of the p53- and pRb-tumor suppressor pathways to the development of most, if not all types of cancer, is unequivocal. However, critical questions remain concerning the relative roles of specific p53- and pRb-pathway members (e.g., ARF and INK4A, respectively) in any given cell type. Using animal- and cell culture-based melanoma models we found that in melanocytes oncogene-induced senescence, a barrier against early tumor progression, can be overcome by a deficiency in ARF, but not p53, facilitating development of melanoma (Ha et al., Proc. Natl. Acad. Sci. 104: 10968-73, 2007). Our data help explain in human melanoma the relative abundance and paucity of mutations in ARF and TP53, respectively, and show that ARF and p53 suppress tumorigenesis through diverse, lineage-dependent mechanisms. In vitro and in vivo models based on these genetically engineered melanocytes have been used to identify novel regulators of differentiation, malignancy and metastasis. For example, we compared the oncogenic roles of the three major NRAS downstream effectors, RAF, PI3K and RAL guanine exchange factor (RalGEF) (Mishra et al., Oncogene 29:2249-2256, 2010). Although no single downstream pathway could recapitulate all the consequences of oncogenic NRAS, we found a prominent role for BRAF and PI3K in melanocyte senescence and invasiveness, respectively. We also discovered that constitutive RalGEF activation stimulated anchorage-independent growth, indicating that this often overlooked pathway should be evaluated as a possible therapeutic target. We have also begun re-exploring the prospects of differentiation therapy. INK4A/ARF-deficient melanocytes transformed by mutant NRAS were used in a high-throughput screen of the LOPAC drug library to identify agents that would induce re-differentiation of malignant melanoma cells, thus converting them to a more benign state. Several interesting drugs were identified, including two clinically active agents that could be re-purposed. Although an extensive accumulation of epidemiological evidence supportsa fundamental role for UV in melanoma, the specific UV-affected molecular pathways and mechanisms remain largely unidentified. Few UV signature mutations have been found in genes thought to contribute to melanoma, although deep sequencing of the melanoma genome revealed a broad spectrum of UVB signature mutations. We have suggested that mechanisms other than UV-induced DNA mutagenesis may also be important in melanoma initiation. To determine the role(s) of UV in melanoma in vivo, we developed an experimental mouse model (iDCT-GFP) that allows melanocytes, specifically and inducibly labeled with green fluorescent protein (GFP), to be highly purified from disaggregated mouse skin by FACS following UV irradiation in vivo. We identified a pattern of UVB induced gene expression changes in melanocytes isolated from mice that are consistent with inflammatory alterations and may spare melanocytes post-UV remodeling-associated destruction. We have identified an interferon (IFN)-gamma signaling signature arising in melanocytes after neonatal UV irradiation. The source was macrophages recruited to the skin after UV exposure; IFN-gamma in turn activated melanocytes and the expression of genes that could facilitate immunoevasion. Transplanted neonatal macrophages were found to significantly enhance melanoma growth in vivo in an IFN-gamma-dependent fashion. This was surprising considering that IFN-alpha has been used to treat melanoma patients, albeit with limited success. We hypothesized that melanomas escape immune destruction by co-opting these pathways already hard-wired in melanocytes, and suggested that the IFN-gamma signaling pathway may represent a promising therapeutic target for melanoma patients (Zaidi et al., Nature 469:548-553, 2011). To address the UV mutation question we are also subjecting in vivo-exposed melanocytes to RNA sequencing. We will use fluorescence activated cell sorting to isolate GFP-labeled melanocytes from all stages of melanoma development relevant to human disease in an attempt to catalog their precise genomic alterations. We anticipate that this in vivo model will provide novel insights into the nature of UV-induced damage, and the mechanisms by which UV provokes melanoma. We have also hypothesized that late stage melanoma cells can co-opt pathways hard-wired into normal developing melanoblasts to achieve a more aggressive and metastatic phenotype. Both the embryonic melanoblast and the metastatic melanoma cell must undergo a similar EMT and become invasive, highly migratory, and survive to colonize at a remote sites. We again employed our iDCT-GFP model to isolate embryonic melanoblasts from key stages of melanocyte development. RNA sequencing and microarray-based gene expression profiling have been performed from representative developmental stages. Genes have been identified whose expression is characteristically up-regulated (or down-regulated) in both melanoblasts and metastatic melanoma relative to adult melanocytes, which may represent new therapeutic targets against metastasis in melanoma. Validating genes that have come out of this screen to date include Twist1/2, Osteopontin, MMP2/9 and Tenascin C. Novel candidates, which include genes that regulate protein turnover and trafficing, metabolism, and neural developoment are currently being evaluated for a role in metastasis using siRNA knockdown in metastatic human melanoma cells and tail vein injections. Several candidates have now been shown to regulate metastatic behavior, and to correlate with melanoma patient survival.